Kaldi单步完美运行AIShell v1 S5之三：三音tri1,tri2,tri3,tri4,tri5

致谢

感谢AIShell在商业化道路上的探索。期待着v3的到来。

机器配置

sv@HP:~$ sudo lsb_release -a
Distributor ID:	Ubuntu
Description:	Ubuntu 18.04.1 LTS
Release:	18.04
Codename:	bionic

sv@HP:~$ cat /proc/cpuinfo | grep model\ name
model name	: Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz
model name	: Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz
model name	: Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz
model name	: Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz
model name	: Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz
model name	: Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz
model name	: Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz
model name	: Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz
model name	: Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz
model name	: Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz
model name	: Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz
model name	: Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz
sv@HP:~$ cat /proc/meminfo | grep MemTotal
MemTotal:       16321360 kB
sv@HP:~$ lspci | grep 'VGA'
01:00.0 VGA compatible controller: NVIDIA Corporation GP104 [GeForce GTX 1070] (rev a1)

Kaldi下AIShell v1详细输出之三：三音triphone

一网打尽。

第五部分：三音结果更新

先看结果。tri4a的时候，还没上DNN之前，错词率已经从Mono的37%降到14%了。

sv@HP:~/lkaldi/egs/aishell/s5$ for x in exp/ */decode_test; do [ -d $x ] && grep WER $x/cer_* | utils/best_wer.sh; done 2>/dev/null

%WER 36.59 [ 38335 / 104765, 849 ins, 3183 del, 34303 sub ] exp/mono/decode_test/cer_10_0.0
%WER 18.83 [ 19727 / 104765, 971 ins, 1161 del, 17595 sub ] exp/tri1/decode_test/cer_13_0.5
%WER 18.79 [ 19684 / 104765, 957 ins, 1142 del, 17585 sub ] exp/tri2/decode_test/cer_14_0.5
%WER 16.84 [ 17643 / 104765, 791 ins, 991 del, 15861 sub ] exp/tri3a/decode_test/cer_14_0.5
%WER 13.63 [ 14277 / 104765, 762 ins, 639 del, 12876 sub ] exp/tri4a/decode_test/cer_13_0.5
%WER 12.02 [ 12594 / 104765, 745 ins, 496 del, 11353 sub ] exp/tri5a/decode_test/cer_13_0.5

第六部分：三音Tri1训练、解码、校准

sv@HP:~/lkaldi/egs/aishell/s5$ # train tri1 [first triphone pass]
sv@HP:~/lkaldi/egs/aishell/s5$ steps/train_deltas.sh --cmd "$train_cmd" 2500 20000 data/train data/lang exp/mono_ali exp/tri1 || exit 1;
steps/train_deltas.sh --cmd run.pl --mem 8G 2500 20000 data/train data/lang exp/mono_ali exp/tri1
steps/train_deltas.sh: accumulating tree stats
steps/train_deltas.sh: getting questions for tree-building, via clustering
steps/train_deltas.sh: building the tree
steps/train_deltas.sh: converting alignments from exp/mono_ali to use current tree
steps/train_deltas.sh: compiling graphs of transcripts
steps/train_deltas.sh: training pass 1
steps/train_deltas.sh: training pass 2
steps/train_deltas.sh: training pass 3
steps/train_deltas.sh: training pass 4
steps/train_deltas.sh: training pass 5
steps/train_deltas.sh: training pass 6
steps/train_deltas.sh: training pass 7
steps/train_deltas.sh: training pass 8
steps/train_deltas.sh: training pass 9
steps/train_deltas.sh: training pass 10
steps/train_deltas.sh: aligning data
steps/train_deltas.sh: training pass 11
steps/train_deltas.sh: training pass 12
steps/train_deltas.sh: training pass 13
steps/train_deltas.sh: training pass 14
steps/train_deltas.sh: training pass 15
steps/train_deltas.sh: training pass 16
steps/train_deltas.sh: training pass 17
steps/train_deltas.sh: training pass 18
steps/train_deltas.sh: training pass 19
steps/train_deltas.sh: training pass 20
steps/train_deltas.sh: aligning data
steps/train_deltas.sh: training pass 21
steps/train_deltas.sh: training pass 22
steps/train_deltas.sh: training pass 23
steps/train_deltas.sh: training pass 24
steps/train_deltas.sh: training pass 25
steps/train_deltas.sh: training pass 26
steps/train_deltas.sh: training pass 27
steps/train_deltas.sh: training pass 28
steps/train_deltas.sh: training pass 29
steps/train_deltas.sh: training pass 30
steps/train_deltas.sh: aligning data
steps/train_deltas.sh: training pass 31
steps/train_deltas.sh: training pass 32
steps/train_deltas.sh: training pass 33
steps/train_deltas.sh: training pass 34
steps/diagnostic/analyze_alignments.sh --cmd run.pl --mem 8G data/lang exp/tri1
steps/diagnostic/analyze_alignments.sh: see stats in exp/tri1/log/analyze_alignments.log
877 warnings in exp/tri1/log/acc.*.*.log
1 warnings in exp/tri1/log/build_tree.log
1 warnings in exp/tri1/log/compile_questions.log
3146 warnings in exp/tri1/log/align.*.*.log
exp/tri1: nj=10 align prob=-79.45 over 150.17h [retry=0.5%, fail=0.0%] states=2080 gauss=20046 tree-impr=4.49
steps/train_deltas.sh: Done training system with delta+delta-delta features in exp/tri1
sv@HP:~/lkaldi/egs/aishell/s5$ 
sv@HP:~/lkaldi/egs/aishell/s5$ # decode tri1
sv@HP:~/lkaldi/egs/aishell/s5$ utils/mkgraph.sh data/lang_test exp/tri1 exp/tri1/graph || exit 1;
tree-info exp/tri1/tree 
tree-info exp/tri1/tree 
fstcomposecontext --context-size=3 --central-position=1 --read-disambig-syms=data/lang_test/phones/disambig.int --write-disambig-syms=data/lang_test/tmp/disambig_ilabels_3_1.int data/lang_test/tmp/ilabels_3_1.10620 data/lang_test/tmp/LG.fst 
fstisstochastic data/lang_test/tmp/CLG_3_1.fst 
0 -0.0666824
[info]: CLG not stochastic.
make-h-transducer --disambig-syms-out=exp/tri1/graph/disambig_tid.int --transition-scale=1.0 data/lang_test/tmp/ilabels_3_1 exp/tri1/tree exp/tri1/final.mdl 
fstrmepslocal 
fsttablecompose exp/tri1/graph/Ha.fst data/lang_test/tmp/CLG_3_1.fst 
fstrmsymbols exp/tri1/graph/disambig_tid.int 
fstminimizeencoded 
fstdeterminizestar --use-log=true 
fstisstochastic exp/tri1/graph/HCLGa.fst 
0.000487832 -0.178947
HCLGa is not stochastic
add-self-loops --self-loop-scale=0.1 --reorder=true exp/tri1/final.mdl exp/tri1/graph/HCLGa.fst 
sv@HP:~/lkaldi/egs/aishell/s5$ steps/decode.sh --cmd "$decode_cmd" --config conf/decode.config --nj 10 exp/tri1/graph data/dev exp/tri1/decode_dev
steps/decode.sh --cmd run.pl --mem 8G --config conf/decode.config --nj 10 exp/tri1/graph data/dev exp/tri1/decode_dev
decode.sh: feature type is delta
steps/diagnostic/analyze_lats.sh --cmd run.pl --mem 8G exp/tri1/graph exp/tri1/decode_dev
steps/diagnostic/analyze_lats.sh: see stats in exp/tri1/decode_dev/log/analyze_alignments.log
Overall, lattice depth (10,50,90-percentile)=(1,4,30) and mean=12.1
steps/diagnostic/analyze_lats.sh: see stats in exp/tri1/decode_dev/log/analyze_lattice_depth_stats.log
+ steps/score_kaldi.sh --cmd 'run.pl --mem 8G' data/dev exp/tri1/graph exp/tri1/decode_dev
steps/score_kaldi.sh --cmd run.pl --mem 8G data/dev exp/tri1/graph exp/tri1/decode_dev
steps/score_kaldi.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ steps/scoring/score_kaldi_cer.sh --stage 2 --cmd 'run.pl --mem 8G' data/dev exp/tri1/graph exp/tri1/decode_dev
steps/scoring/score_kaldi_cer.sh --stage 2 --cmd run.pl --mem 8G data/dev exp/tri1/graph exp/tri1/decode_dev
steps/scoring/score_kaldi_cer.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ echo 'local/score.sh: Done'
local/score.sh: Done
sv@HP:~/lkaldi/egs/aishell/s5$ steps/decode.sh --cmd "$decode_cmd" --config conf/decode.config --nj 10 exp/tri1/graph data/test exp/tri1/decode_test
steps/decode.sh --cmd run.pl --mem 8G --config conf/decode.config --nj 10 exp/tri1/graph data/test exp/tri1/decode_test
decode.sh: feature type is delta
steps/diagnostic/analyze_lats.sh --cmd run.pl --mem 8G exp/tri1/graph exp/tri1/decode_test
steps/diagnostic/analyze_lats.sh: see stats in exp/tri1/decode_test/log/analyze_alignments.log
Overall, lattice depth (10,50,90-percentile)=(1,4,40) and mean=15.4
steps/diagnostic/analyze_lats.sh: see stats in exp/tri1/decode_test/log/analyze_lattice_depth_stats.log
+ steps/score_kaldi.sh --cmd 'run.pl --mem 8G' data/test exp/tri1/graph exp/tri1/decode_test
steps/score_kaldi.sh --cmd run.pl --mem 8G data/test exp/tri1/graph exp/tri1/decode_test
steps/score_kaldi.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ steps/scoring/score_kaldi_cer.sh --stage 2 --cmd 'run.pl --mem 8G' data/test exp/tri1/graph exp/tri1/decode_test
steps/scoring/score_kaldi_cer.sh --stage 2 --cmd run.pl --mem 8G data/test exp/tri1/graph exp/tri1/decode_test
steps/scoring/score_kaldi_cer.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ echo 'local/score.sh: Done'
local/score.sh: Done

sv@HP:~/lkaldi/egs/aishell/s5$ 
sv@HP:~/lkaldi/egs/aishell/s5$ # align tri1
sv@HP:~/lkaldi/egs/aishell/s5$ steps/align_si.sh --cmd "$train_cmd" --nj 10 data/train data/lang exp/tri1 exp/tri1_ali || exit 1;
steps/align_si.sh --cmd run.pl --mem 8G --nj 10 data/train data/lang exp/tri1 exp/tri1_ali
steps/align_si.sh: feature type is delta
steps/align_si.sh: aligning data in data/train using model from exp/tri1, putting alignments in exp/tri1_ali
steps/diagnostic/analyze_alignments.sh --cmd run.pl --mem 8G data/lang exp/tri1_ali
steps/diagnostic/analyze_alignments.sh: see stats in exp/tri1_ali/log/analyze_alignments.log
steps/align_si.sh: done aligning data.

第七部分：三音素tri2（delta+delta-deltas）

sv@HP:~/lkaldi/egs/aishell/s5$ # train tri2 [delta+delta-deltas]
sv@HP:~/lkaldi/egs/aishell/s5$ steps/train_deltas.sh --cmd "$train_cmd" 2500 20000 data/train data/lang exp/tri1_ali exp/tri2 || exit 1;
steps/train_deltas.sh --cmd run.pl --mem 8G 2500 20000 data/train data/lang exp/tri1_ali exp/tri2
steps/train_deltas.sh: accumulating tree stats
steps/train_deltas.sh: getting questions for tree-building, via clustering
steps/train_deltas.sh: building the tree
steps/train_deltas.sh: converting alignments from exp/tri1_ali to use current tree
steps/train_deltas.sh: compiling graphs of transcripts
steps/train_deltas.sh: training pass 1
steps/train_deltas.sh: training pass 2
steps/train_deltas.sh: training pass 3
steps/train_deltas.sh: training pass 4
steps/train_deltas.sh: training pass 5
steps/train_deltas.sh: training pass 6
steps/train_deltas.sh: training pass 7
steps/train_deltas.sh: training pass 8
steps/train_deltas.sh: training pass 9
steps/train_deltas.sh: training pass 10
steps/train_deltas.sh: aligning data
steps/train_deltas.sh: training pass 11
steps/train_deltas.sh: training pass 12
steps/train_deltas.sh: training pass 13
steps/train_deltas.sh: training pass 14
steps/train_deltas.sh: training pass 15
steps/train_deltas.sh: training pass 16
steps/train_deltas.sh: training pass 17
steps/train_deltas.sh: training pass 18
steps/train_deltas.sh: training pass 19
steps/train_deltas.sh: training pass 20
steps/train_deltas.sh: aligning data
steps/train_deltas.sh: training pass 21
steps/train_deltas.sh: training pass 22
steps/train_deltas.sh: training pass 23
steps/train_deltas.sh: training pass 24
steps/train_deltas.sh: training pass 25
steps/train_deltas.sh: training pass 26
steps/train_deltas.sh: training pass 27
steps/train_deltas.sh: training pass 28
steps/train_deltas.sh: training pass 29
steps/train_deltas.sh: training pass 30
steps/train_deltas.sh: aligning data
steps/train_deltas.sh: training pass 31
steps/train_deltas.sh: training pass 32
steps/train_deltas.sh: training pass 33
steps/train_deltas.sh: training pass 34
steps/diagnostic/analyze_alignments.sh --cmd run.pl --mem 8G data/lang exp/tri2
steps/diagnostic/analyze_alignments.sh: see stats in exp/tri2/log/analyze_alignments.log
1 warnings in exp/tri2/log/compile_questions.log
2213 warnings in exp/tri2/log/align.*.*.log
1 warnings in exp/tri2/log/build_tree.log
560 warnings in exp/tri2/log/acc.*.*.log
exp/tri2: nj=10 align prob=-79.41 over 150.18h [retry=0.4%, fail=0.0%] states=2088 gauss=20034 tree-impr=4.82
steps/train_deltas.sh: Done training system with delta+delta-delta features in exp/tri2
sv@HP:~/lkaldi/egs/aishell/s5$ 
sv@HP:~/lkaldi/egs/aishell/s5$ # decode tri2
sv@HP:~/lkaldi/egs/aishell/s5$ utils/mkgraph.sh data/lang_test exp/tri2 exp/tri2/graph
tree-info exp/tri2/tree 
tree-info exp/tri2/tree 
make-h-transducer --disambig-syms-out=exp/tri2/graph/disambig_tid.int --transition-scale=1.0 data/lang_test/tmp/ilabels_3_1 exp/tri2/tree exp/tri2/final.mdl 
fsttablecompose exp/tri2/graph/Ha.fst data/lang_test/tmp/CLG_3_1.fst 
fstdeterminizestar --use-log=true 
fstrmsymbols exp/tri2/graph/disambig_tid.int 
fstrmepslocal 
fstminimizeencoded 
fstisstochastic exp/tri2/graph/HCLGa.fst 
0.000487022 -0.178947
HCLGa is not stochastic
add-self-loops --self-loop-scale=0.1 --reorder=true exp/tri2/final.mdl exp/tri2/graph/HCLGa.fst 
sv@HP:~/lkaldi/egs/aishell/s5$ steps/decode.sh --cmd "$decode_cmd" --config conf/decode.config --nj 10 exp/tri2/graph data/dev exp/tri2/decode_dev
steps/decode.sh --cmd run.pl --mem 8G --config conf/decode.config --nj 10 exp/tri2/graph data/dev exp/tri2/decode_dev
decode.sh: feature type is delta
steps/diagnostic/analyze_lats.sh --cmd run.pl --mem 8G exp/tri2/graph exp/tri2/decode_dev
steps/diagnostic/analyze_lats.sh: see stats in exp/tri2/decode_dev/log/analyze_alignments.log
Overall, lattice depth (10,50,90-percentile)=(1,4,30) and mean=11.7
steps/diagnostic/analyze_lats.sh: see stats in exp/tri2/decode_dev/log/analyze_lattice_depth_stats.log
+ steps/score_kaldi.sh --cmd 'run.pl --mem 8G' data/dev exp/tri2/graph exp/tri2/decode_dev
steps/score_kaldi.sh --cmd run.pl --mem 8G data/dev exp/tri2/graph exp/tri2/decode_dev
steps/score_kaldi.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ steps/scoring/score_kaldi_cer.sh --stage 2 --cmd 'run.pl --mem 8G' data/dev exp/tri2/graph exp/tri2/decode_dev
steps/scoring/score_kaldi_cer.sh --stage 2 --cmd run.pl --mem 8G data/dev exp/tri2/graph exp/tri2/decode_dev
steps/scoring/score_kaldi_cer.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ echo 'local/score.sh: Done'
local/score.sh: Done
sv@HP:~/lkaldi/egs/aishell/s5$ steps/decode.sh --cmd "$decode_cmd" --config conf/decode.config --nj 10 exp/tri2/graph data/test exp/tri2/decode_test
steps/decode.sh --cmd run.pl --mem 8G --config conf/decode.config --nj 10 exp/tri2/graph data/test exp/tri2/decode_test
decode.sh: feature type is delta
steps/diagnostic/analyze_lats.sh --cmd run.pl --mem 8G exp/tri2/graph exp/tri2/decode_test
steps/diagnostic/analyze_lats.sh: see stats in exp/tri2/decode_test/log/analyze_alignments.log
Overall, lattice depth (10,50,90-percentile)=(1,4,39) and mean=15.2
steps/diagnostic/analyze_lats.sh: see stats in exp/tri2/decode_test/log/analyze_lattice_depth_stats.log
+ steps/score_kaldi.sh --cmd 'run.pl --mem 8G' data/test exp/tri2/graph exp/tri2/decode_test
steps/score_kaldi.sh --cmd run.pl --mem 8G data/test exp/tri2/graph exp/tri2/decode_test
steps/score_kaldi.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ steps/scoring/score_kaldi_cer.sh --stage 2 --cmd 'run.pl --mem 8G' data/test exp/tri2/graph exp/tri2/decode_test
steps/scoring/score_kaldi_cer.sh --stage 2 --cmd run.pl --mem 8G data/test exp/tri2/graph exp/tri2/decode_test
steps/scoring/score_kaldi_cer.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ echo 'local/score.sh: Done'
local/score.sh: Done

第八部分：三音素tri3a[LDA+MLLT]

sv@HP:~/lkaldi/egs/aishell/s5$ # train and decode tri2b [LDA+MLLT]
sv@HP:~/lkaldi/egs/aishell/s5$ steps/align_si.sh --cmd "$train_cmd" --nj 10 data/train data/lang exp/tri2 exp/tri2_ali || exit 1;
steps/align_si.sh --cmd run.pl --mem 8G --nj 10 data/train data/lang exp/tri2 exp/tri2_ali
steps/align_si.sh: feature type is delta
steps/align_si.sh: aligning data in data/train using model from exp/tri2, putting alignments in exp/tri2_ali
steps/diagnostic/analyze_alignments.sh --cmd run.pl --mem 8G data/lang exp/tri2_ali
steps/diagnostic/analyze_alignments.sh: see stats in exp/tri2_ali/log/analyze_alignments.log
steps/align_si.sh: done aligning data.
sv@HP:~/lkaldi/egs/aishell/s5$ 
sv@HP:~/lkaldi/egs/aishell/s5$ # Train tri3a, which is LDA+MLLT,
sv@HP:~/lkaldi/egs/aishell/s5$ steps/train_lda_mllt.sh --cmd "$train_cmd" 2500 20000 data/train data/lang exp/tri2_ali exp/tri3a || exit 1;
steps/train_lda_mllt.sh --cmd run.pl --mem 8G 2500 20000 data/train data/lang exp/tri2_ali exp/tri3a
steps/train_lda_mllt.sh: Accumulating LDA statistics.
steps/train_lda_mllt.sh: Accumulating tree stats
steps/train_lda_mllt.sh: Getting questions for tree clustering.
steps/train_lda_mllt.sh: Building the tree
steps/train_lda_mllt.sh: Initializing the model
steps/train_lda_mllt.sh: Converting alignments from exp/tri2_ali to use current tree
steps/train_lda_mllt.sh: Compiling graphs of transcripts
Training pass 1
Training pass 2
steps/train_lda_mllt.sh: Estimating MLLT
Training pass 3
Training pass 4
steps/train_lda_mllt.sh: Estimating MLLT
Training pass 5
Training pass 6
steps/train_lda_mllt.sh: Estimating MLLT
Training pass 7
Training pass 8
Training pass 9
Training pass 10
Aligning data
Training pass 11
Training pass 12
steps/train_lda_mllt.sh: Estimating MLLT
Training pass 13
Training pass 14
Training pass 15
Training pass 16
Training pass 17
Training pass 18
Training pass 19
Training pass 20
Aligning data
Training pass 21
Training pass 22
Training pass 23
Training pass 24
Training pass 25
Training pass 26
Training pass 27
Training pass 28
Training pass 29
Training pass 30
Aligning data
Training pass 31
Training pass 32
Training pass 33
Training pass 34
steps/diagnostic/analyze_alignments.sh --cmd run.pl --mem 8G data/lang exp/tri3a
steps/diagnostic/analyze_alignments.sh: see stats in exp/tri3a/log/analyze_alignments.log
4 warnings in exp/tri3a/log/lda_acc.*.log
1 warnings in exp/tri3a/log/build_tree.log
1 warnings in exp/tri3a/log/compile_questions.log
1206 warnings in exp/tri3a/log/align.*.*.log
246 warnings in exp/tri3a/log/acc.*.*.log
exp/tri3a: nj=10 align prob=-48.78 over 150.18h [retry=0.2%, fail=0.0%] states=2112 gauss=20039 tree-impr=5.06 lda-sum=24.62 mllt:impr,logdet=0.94,1.41
steps/train_lda_mllt.sh: Done training system with LDA+MLLT features in exp/tri3a
sv@HP:~/lkaldi/egs/aishell/s5$ 
sv@HP:~/lkaldi/egs/aishell/s5$ utils/mkgraph.sh data/lang_test exp/tri3a exp/tri3a/graph || exit 1;
tree-info exp/tri3a/tree 
tree-info exp/tri3a/tree 
make-h-transducer --disambig-syms-out=exp/tri3a/graph/disambig_tid.int --transition-scale=1.0 data/lang_test/tmp/ilabels_3_1 exp/tri3a/tree exp/tri3a/final.mdl 
fstrmepslocal 
fsttablecompose exp/tri3a/graph/Ha.fst data/lang_test/tmp/CLG_3_1.fst 
fstrmsymbols exp/tri3a/graph/disambig_tid.int 
fstminimizeencoded 
fstdeterminizestar --use-log=true 
fstisstochastic exp/tri3a/graph/HCLGa.fst 
0.000487099 -0.178947
HCLGa is not stochastic
add-self-loops --self-loop-scale=0.1 --reorder=true exp/tri3a/final.mdl exp/tri3a/graph/HCLGa.fst 
sv@HP:~/lkaldi/egs/aishell/s5$ steps/decode.sh --cmd "$decode_cmd" --nj 10 --config conf/decode.config exp/tri3a/graph data/dev exp/tri3a/decode_dev
steps/decode.sh --cmd run.pl --mem 8G --nj 10 --config conf/decode.config exp/tri3a/graph data/dev exp/tri3a/decode_dev
decode.sh: feature type is lda
steps/diagnostic/analyze_lats.sh --cmd run.pl --mem 8G exp/tri3a/graph exp/tri3a/decode_dev
steps/diagnostic/analyze_lats.sh: see stats in exp/tri3a/decode_dev/log/analyze_alignments.log
Overall, lattice depth (10,50,90-percentile)=(1,3,24) and mean=9.6
steps/diagnostic/analyze_lats.sh: see stats in exp/tri3a/decode_dev/log/analyze_lattice_depth_stats.log
+ steps/score_kaldi.sh --cmd 'run.pl --mem 8G' data/dev exp/tri3a/graph exp/tri3a/decode_dev
steps/score_kaldi.sh --cmd run.pl --mem 8G data/dev exp/tri3a/graph exp/tri3a/decode_dev
steps/score_kaldi.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ steps/scoring/score_kaldi_cer.sh --stage 2 --cmd 'run.pl --mem 8G' data/dev exp/tri3a/graph exp/tri3a/decode_dev
steps/scoring/score_kaldi_cer.sh --stage 2 --cmd run.pl --mem 8G data/dev exp/tri3a/graph exp/tri3a/decode_dev
steps/scoring/score_kaldi_cer.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ echo 'local/score.sh: Done'
local/score.sh: Done
sv@HP:~/lkaldi/egs/aishell/s5$ steps/decode.sh --cmd "$decode_cmd" --nj 10 --config conf/decode.config exp/tri3a/graph data/test exp/tri3a/decode_test
steps/decode.sh --cmd run.pl --mem 8G --nj 10 --config conf/decode.config exp/tri3a/graph data/test exp/tri3a/decode_test
decode.sh: feature type is lda
steps/diagnostic/analyze_lats.sh --cmd run.pl --mem 8G exp/tri3a/graph exp/tri3a/decode_test
steps/diagnostic/analyze_lats.sh: see stats in exp/tri3a/decode_test/log/analyze_alignments.log
Overall, lattice depth (10,50,90-percentile)=(1,4,30) and mean=11.9
steps/diagnostic/analyze_lats.sh: see stats in exp/tri3a/decode_test/log/analyze_lattice_depth_stats.log
+ steps/score_kaldi.sh --cmd 'run.pl --mem 8G' data/test exp/tri3a/graph exp/tri3a/decode_test
steps/score_kaldi.sh --cmd run.pl --mem 8G data/test exp/tri3a/graph exp/tri3a/decode_test
steps/score_kaldi.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ steps/scoring/score_kaldi_cer.sh --stage 2 --cmd 'run.pl --mem 8G' data/test exp/tri3a/graph exp/tri3a/decode_test
steps/scoring/score_kaldi_cer.sh --stage 2 --cmd run.pl --mem 8G data/test exp/tri3a/graph exp/tri3a/decode_test
steps/scoring/score_kaldi_cer.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ echo 'local/score.sh: Done'
local/score.sh: Done
sv@HP:~/lkaldi/egs/aishell/s5$ 
sv@HP:~/lkaldi/egs/aishell/s5$ 
sv@HP:~/lkaldi/egs/aishell/s5$ steps/align_fmllr.sh --cmd "$train_cmd" --nj 10 data/train data/lang exp/tri3a exp/tri3a_ali || exit 1;
steps/align_fmllr.sh --cmd run.pl --mem 8G --nj 10 data/train data/lang exp/tri3a exp/tri3a_ali
steps/align_fmllr.sh: feature type is lda
steps/align_fmllr.sh: compiling training graphs
steps/align_fmllr.sh: aligning data in data/train using exp/tri3a/final.mdl and speaker-independent features.
steps/align_fmllr.sh: computing fMLLR transforms
steps/align_fmllr.sh: doing final alignment.
steps/align_fmllr.sh: done aligning data.
steps/diagnostic/analyze_alignments.sh --cmd run.pl --mem 8G data/lang exp/tri3a_ali
steps/diagnostic/analyze_alignments.sh: see stats in exp/tri3a_ali/log/analyze_alignments.log
3 warnings in exp/tri3a_ali/log/fmllr.*.log
231 warnings in exp/tri3a_ali/log/align_pass1.*.log
224 warnings in exp/tri3a_ali/log/align_pass2.*.log

第九部分：三音素tri4

sv@HP:~/lkaldi/egs/aishell/s5$ steps/train_sat.sh --cmd "$train_cmd" 2500 20000 data/train data/lang exp/tri3a_ali exp/tri4a || exit 1;
steps/train_sat.sh --cmd run.pl --mem 8G 2500 20000 data/train data/lang exp/tri3a_ali exp/tri4a
steps/train_sat.sh: feature type is lda
steps/train_sat.sh: Using transforms from exp/tri3a_ali
steps/train_sat.sh: Accumulating tree stats
steps/train_sat.sh: Getting questions for tree clustering.
steps/train_sat.sh: Building the tree
steps/train_sat.sh: Initializing the model
steps/train_sat.sh: Converting alignments from exp/tri3a_ali to use current tree
steps/train_sat.sh: Compiling graphs of transcripts
Pass 1
Pass 2
Estimating fMLLR transforms
Pass 3
Pass 4
Estimating fMLLR transforms
Pass 5
Pass 6
Estimating fMLLR transforms
Pass 7
Pass 8
Pass 9
Pass 10
Aligning data
Pass 11
Pass 12
Estimating fMLLR transforms
Pass 13
Pass 14
Pass 15
Pass 16
Pass 17
Pass 18
Pass 19
Pass 20
Aligning data
Pass 21
Pass 22
Pass 23
Pass 24
Pass 25
Pass 26
Pass 27
Pass 28
Pass 29
Pass 30
Aligning data
Pass 31
Pass 32
Pass 33
Pass 34
steps/diagnostic/analyze_alignments.sh --cmd run.pl --mem 8G data/lang exp/tri4a
steps/diagnostic/analyze_alignments.sh: see stats in exp/tri4a/log/analyze_alignments.log
1 warnings in exp/tri4a/log/build_tree.log
1 warnings in exp/tri4a/log/compile_questions.log
1478 warnings in exp/tri4a/log/align.*.*.log
507 warnings in exp/tri4a/log/acc.*.*.log
29 warnings in exp/tri4a/log/fmllr.*.*.log
steps/train_sat.sh: Likelihood evolution:
-49.3274 -49.1352 -49.013 -48.7941 -48.3565 -47.938 -47.6176 -47.3933 -47.2217 -46.8437 -46.692 -46.4815 -46.3709 -46.2882 -46.2175 -46.1497 -46.0864 -46.0267 -45.9704 -45.8353 -45.7686 -45.7266 -45.6894 -45.6555 -45.6239 -45.5935 -45.5641 -45.5345 -45.5049 -45.4336 -45.3936 -45.373 -45.3598 -45.3509 
exp/tri4a: nj=10 align prob=-48.30 over 150.17h [retry=0.3%, fail=0.0%] states=2112 gauss=20024 fmllr-impr=0.63 over 115.53h tree-impr=7.03
steps/train_sat.sh: done training SAT system in exp/tri4a
sv@HP:~/lkaldi/egs/aishell/s5$ 
sv@HP:~/lkaldi/egs/aishell/s5$ utils/mkgraph.sh data/lang_test exp/tri4a exp/tri4a/graph
tree-info exp/tri4a/tree 
tree-info exp/tri4a/tree 
make-h-transducer --disambig-syms-out=exp/tri4a/graph/disambig_tid.int --transition-scale=1.0 data/lang_test/tmp/ilabels_3_1 exp/tri4a/tree exp/tri4a/final.mdl 
fstrmepslocal 
fsttablecompose exp/tri4a/graph/Ha.fst data/lang_test/tmp/CLG_3_1.fst 
fstrmsymbols exp/tri4a/graph/disambig_tid.int 
fstminimizeencoded 
fstdeterminizestar --use-log=true 
fstisstochastic exp/tri4a/graph/HCLGa.fst 
0.000487832 -0.178947
HCLGa is not stochastic
add-self-loops --self-loop-scale=0.1 --reorder=true exp/tri4a/final.mdl exp/tri4a/graph/HCLGa.fst 
sv@HP:~/lkaldi/egs/aishell/s5$ steps/decode_fmllr.sh --cmd "$decode_cmd" --nj 10 --config conf/decode.config exp/tri4a/graph data/dev exp/tri4a/decode_dev
steps/decode_fmllr.sh --cmd run.pl --mem 8G --nj 10 --config conf/decode.config exp/tri4a/graph data/dev exp/tri4a/decode_dev
steps/decode.sh --scoring-opts  --num-threads 1 --skip-scoring false --acwt 0.083333 --nj 10 --cmd run.pl --mem 8G --beam 8.0 --model exp/tri4a/final.alimdl --max-active 2000 exp/tri4a/graph data/dev exp/tri4a/decode_dev.si
decode.sh: feature type is lda
steps/diagnostic/analyze_lats.sh --cmd run.pl --mem 8G exp/tri4a/graph exp/tri4a/decode_dev.si
steps/diagnostic/analyze_lats.sh: see stats in exp/tri4a/decode_dev.si/log/analyze_alignments.log
Overall, lattice depth (10,50,90-percentile)=(1,3,16) and mean=6.7
steps/diagnostic/analyze_lats.sh: see stats in exp/tri4a/decode_dev.si/log/analyze_lattice_depth_stats.log
+ steps/score_kaldi.sh --cmd 'run.pl --mem 8G' data/dev exp/tri4a/graph exp/tri4a/decode_dev.si
steps/score_kaldi.sh --cmd run.pl --mem 8G data/dev exp/tri4a/graph exp/tri4a/decode_dev.si
steps/score_kaldi.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ steps/scoring/score_kaldi_cer.sh --stage 2 --cmd 'run.pl --mem 8G' data/dev exp/tri4a/graph exp/tri4a/decode_dev.si
steps/scoring/score_kaldi_cer.sh --stage 2 --cmd run.pl --mem 8G data/dev exp/tri4a/graph exp/tri4a/decode_dev.si
steps/scoring/score_kaldi_cer.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ echo 'local/score.sh: Done'
local/score.sh: Done
steps/decode_fmllr.sh: feature type is lda
steps/decode_fmllr.sh: getting first-pass fMLLR transforms.
steps/decode_fmllr.sh: doing main lattice generation phase
steps/decode_fmllr.sh: estimating fMLLR transforms a second time.
steps/decode_fmllr.sh: doing a final pass of acoustic rescoring.
steps/diagnostic/analyze_lats.sh --cmd run.pl --mem 8G exp/tri4a/graph exp/tri4a/decode_dev
steps/diagnostic/analyze_lats.sh: see stats in exp/tri4a/decode_dev/log/analyze_alignments.log
Overall, lattice depth (10,50,90-percentile)=(1,3,17) and mean=7.0
steps/diagnostic/analyze_lats.sh: see stats in exp/tri4a/decode_dev/log/analyze_lattice_depth_stats.log
+ steps/score_kaldi.sh --cmd 'run.pl --mem 8G' data/dev exp/tri4a/graph exp/tri4a/decode_dev
steps/score_kaldi.sh --cmd run.pl --mem 8G data/dev exp/tri4a/graph exp/tri4a/decode_dev
steps/score_kaldi.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ steps/scoring/score_kaldi_cer.sh --stage 2 --cmd 'run.pl --mem 8G' data/dev exp/tri4a/graph exp/tri4a/decode_dev
steps/scoring/score_kaldi_cer.sh --stage 2 --cmd run.pl --mem 8G data/dev exp/tri4a/graph exp/tri4a/decode_dev
steps/scoring/score_kaldi_cer.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ echo 'local/score.sh: Done'
local/score.sh: Done
sv@HP:~/lkaldi/egs/aishell/s5$ steps/decode_fmllr.sh --cmd "$decode_cmd" --nj 10 --config conf/decode.config exp/tri4a/graph data/test exp/tri4a/decode_test
steps/decode_fmllr.sh --cmd run.pl --mem 8G --nj 10 --config conf/decode.config exp/tri4a/graph data/test exp/tri4a/decode_test
steps/decode.sh --scoring-opts  --num-threads 1 --skip-scoring false --acwt 0.083333 --nj 10 --cmd run.pl --mem 8G --beam 8.0 --model exp/tri4a/final.alimdl --max-active 2000 exp/tri4a/graph data/test exp/tri4a/decode_test.si
decode.sh: feature type is lda
steps/diagnostic/analyze_lats.sh --cmd run.pl --mem 8G exp/tri4a/graph exp/tri4a/decode_test.si
steps/diagnostic/analyze_lats.sh: see stats in exp/tri4a/decode_test.si/log/analyze_alignments.log
Overall, lattice depth (10,50,90-percentile)=(1,3,19) and mean=7.9
steps/diagnostic/analyze_lats.sh: see stats in exp/tri4a/decode_test.si/log/analyze_lattice_depth_stats.log
+ steps/score_kaldi.sh --cmd 'run.pl --mem 8G' data/test exp/tri4a/graph exp/tri4a/decode_test.si
steps/score_kaldi.sh --cmd run.pl --mem 8G data/test exp/tri4a/graph exp/tri4a/decode_test.si
steps/score_kaldi.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ steps/scoring/score_kaldi_cer.sh --stage 2 --cmd 'run.pl --mem 8G' data/test exp/tri4a/graph exp/tri4a/decode_test.si
steps/scoring/score_kaldi_cer.sh --stage 2 --cmd run.pl --mem 8G data/test exp/tri4a/graph exp/tri4a/decode_test.si
steps/scoring/score_kaldi_cer.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ echo 'local/score.sh: Done'
local/score.sh: Done
steps/decode_fmllr.sh: feature type is lda
steps/decode_fmllr.sh: getting first-pass fMLLR transforms.
steps/decode_fmllr.sh: doing main lattice generation phase
steps/decode_fmllr.sh: estimating fMLLR transforms a second time.
steps/decode_fmllr.sh: doing a final pass of acoustic rescoring.
steps/diagnostic/analyze_lats.sh --cmd run.pl --mem 8G exp/tri4a/graph exp/tri4a/decode_test
steps/diagnostic/analyze_lats.sh: see stats in exp/tri4a/decode_test/log/analyze_alignments.log
Overall, lattice depth (10,50,90-percentile)=(1,3,20) and mean=8.2
steps/diagnostic/analyze_lats.sh: see stats in exp/tri4a/decode_test/log/analyze_lattice_depth_stats.log
+ steps/score_kaldi.sh --cmd 'run.pl --mem 8G' data/test exp/tri4a/graph exp/tri4a/decode_test
steps/score_kaldi.sh --cmd run.pl --mem 8G data/test exp/tri4a/graph exp/tri4a/decode_test
steps/score_kaldi.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ steps/scoring/score_kaldi_cer.sh --stage 2 --cmd 'run.pl --mem 8G' data/test exp/tri4a/graph exp/tri4a/decode_test
steps/scoring/score_kaldi_cer.sh --stage 2 --cmd run.pl --mem 8G data/test exp/tri4a/graph exp/tri4a/decode_test
steps/scoring/score_kaldi_cer.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ echo 'local/score.sh: Done'
local/score.sh: Done
sv@HP:~/lkaldi/egs/aishell/s5$ 
sv@HP:~/lkaldi/egs/aishell/s5$ steps/align_fmllr.sh  --cmd "$train_cmd" --nj 10 data/train data/lang exp/tri4a exp/tri4a_ali
steps/align_fmllr.sh --cmd run.pl --mem 8G --nj 10 data/train data/lang exp/tri4a exp/tri4a_ali
steps/align_fmllr.sh: feature type is lda
steps/align_fmllr.sh: compiling training graphs
steps/align_fmllr.sh: aligning data in data/train using exp/tri4a/final.alimdl and speaker-independent features.
steps/align_fmllr.sh: computing fMLLR transforms
steps/align_fmllr.sh: doing final alignment.
steps/align_fmllr.sh: done aligning data.
steps/diagnostic/analyze_alignments.sh --cmd run.pl --mem 8G data/lang exp/tri4a_ali
steps/diagnostic/analyze_alignments.sh: see stats in exp/tri4a_ali/log/analyze_alignments.log
330 warnings in exp/tri4a_ali/log/align_pass2.*.log
2 warnings in exp/tri4a_ali/log/fmllr.*.log
194 warnings in exp/tri4a_ali/log/align_pass1.*.log
sv@HP:~/lkaldi/egs/aishell/s5$ 

第十部分：三音素tri5（SAT）

sv@HP:~/lkaldi/egs/aishell/s5$ 
sv@HP:~/lkaldi/egs/aishell/s5$ # Building a larger SAT system.
sv@HP:~/lkaldi/egs/aishell/s5$ 
sv@HP:~/lkaldi/egs/aishell/s5$ steps/train_sat.sh --cmd "$train_cmd" 3500 100000 data/train data/lang exp/tri4a_ali exp/tri5a || exit 1;
steps/train_sat.sh --cmd run.pl --mem 8G 3500 100000 data/train data/lang exp/tri4a_ali exp/tri5a
steps/train_sat.sh: feature type is lda
steps/train_sat.sh: Using transforms from exp/tri4a_ali
steps/train_sat.sh: Accumulating tree stats
steps/train_sat.sh: Getting questions for tree clustering.
steps/train_sat.sh: Building the tree
steps/train_sat.sh: Initializing the model
steps/train_sat.sh: Converting alignments from exp/tri4a_ali to use current tree
steps/train_sat.sh: Compiling graphs of transcripts
Pass 1
Pass 2
Estimating fMLLR transforms
Pass 3
Pass 4
Estimating fMLLR transforms
Pass 5
Pass 6
Estimating fMLLR transforms
Pass 7
Pass 8
Pass 9
Pass 10
Aligning data
Pass 11
Pass 12
Estimating fMLLR transforms
Pass 13
Pass 14
Pass 15
Pass 16
Pass 17
Pass 18
Pass 19
Pass 20
Aligning data
Pass 21
Pass 22
Pass 23
Pass 24
Pass 25
Pass 26
Pass 27
Pass 28
Pass 29
Pass 30
Aligning data
Pass 31
Pass 32
Pass 33
Pass 34
steps/diagnostic/analyze_alignments.sh --cmd run.pl --mem 8G data/lang exp/tri5a
steps/diagnostic/analyze_alignments.sh: see stats in exp/tri5a/log/analyze_alignments.log
29 warnings in exp/tri5a/log/fmllr.*.*.log
732 warnings in exp/tri5a/log/align.*.*.log
219 warnings in exp/tri5a/log/acc.*.*.log
1 warnings in exp/tri5a/log/compile_questions.log
1 warnings in exp/tri5a/log/build_tree.log
steps/train_sat.sh: Likelihood evolution:
-48.6633 -48.7036 -48.6523 -48.4781 -47.9072 -47.1871 -46.6394 -46.285 -46.0195 -45.6945 -45.5263 -45.2563 -45.1266 -45.0326 -44.9479 -44.8699 -44.7987 -44.7348 -44.6769 -44.5585 -44.4903 -44.4448 -44.4045 -44.3665 -44.3303 -44.2956 -44.2623 -44.2301 -44.199 -44.1349 -44.0921 -44.0675 -44.0501 -44.0377 
exp/tri5a: nj=10 align prob=-47.17 over 150.19h [retry=0.1%, fail=0.0%] states=3040 gauss=100108 fmllr-impr=0.26 over 116.61h tree-impr=7.63
steps/train_sat.sh: done training SAT system in exp/tri5a
sv@HP:~/lkaldi/egs/aishell/s5$ 
sv@HP:~/lkaldi/egs/aishell/s5$ utils/mkgraph.sh data/lang_test exp/tri5a exp/tri5a/graph || exit 1;
tree-info exp/tri5a/tree 
tree-info exp/tri5a/tree 
make-h-transducer --disambig-syms-out=exp/tri5a/graph/disambig_tid.int --transition-scale=1.0 data/lang_test/tmp/ilabels_3_1 exp/tri5a/tree exp/tri5a/final.mdl 
fstrmepslocal 
fsttablecompose exp/tri5a/graph/Ha.fst data/lang_test/tmp/CLG_3_1.fst 
fstdeterminizestar --use-log=true 
fstminimizeencoded 
fstrmsymbols exp/tri5a/graph/disambig_tid.int 
fstisstochastic exp/tri5a/graph/HCLGa.fst 
0.00048784 -0.178947
HCLGa is not stochastic
add-self-loops --self-loop-scale=0.1 --reorder=true exp/tri5a/final.mdl exp/tri5a/graph/HCLGa.fst 
sv@HP:~/lkaldi/egs/aishell/s5$ steps/decode_fmllr.sh --cmd "$decode_cmd" --nj 10 --config conf/decode.config exp/tri5a/graph data/dev exp/tri5a/decode_dev || exit 1;
steps/decode_fmllr.sh --cmd run.pl --mem 8G --nj 10 --config conf/decode.config exp/tri5a/graph data/dev exp/tri5a/decode_dev
steps/decode.sh --scoring-opts  --num-threads 1 --skip-scoring false --acwt 0.083333 --nj 10 --cmd run.pl --mem 8G --beam 8.0 --model exp/tri5a/final.alimdl --max-active 2000 exp/tri5a/graph data/dev exp/tri5a/decode_dev.si
decode.sh: feature type is lda
steps/diagnostic/analyze_lats.sh --cmd run.pl --mem 8G exp/tri5a/graph exp/tri5a/decode_dev.si
steps/diagnostic/analyze_lats.sh: see stats in exp/tri5a/decode_dev.si/log/analyze_alignments.log
Overall, lattice depth (10,50,90-percentile)=(1,2,14) and mean=5.7
steps/diagnostic/analyze_lats.sh: see stats in exp/tri5a/decode_dev.si/log/analyze_lattice_depth_stats.log
+ steps/score_kaldi.sh --cmd 'run.pl --mem 8G' data/dev exp/tri5a/graph exp/tri5a/decode_dev.si
steps/score_kaldi.sh --cmd run.pl --mem 8G data/dev exp/tri5a/graph exp/tri5a/decode_dev.si
steps/score_kaldi.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ steps/scoring/score_kaldi_cer.sh --stage 2 --cmd 'run.pl --mem 8G' data/dev exp/tri5a/graph exp/tri5a/decode_dev.si
steps/scoring/score_kaldi_cer.sh --stage 2 --cmd run.pl --mem 8G data/dev exp/tri5a/graph exp/tri5a/decode_dev.si
steps/scoring/score_kaldi_cer.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ echo 'local/score.sh: Done'
local/score.sh: Done
steps/decode_fmllr.sh: feature type is lda
steps/decode_fmllr.sh: getting first-pass fMLLR transforms.
steps/decode_fmllr.sh: doing main lattice generation phase
steps/decode_fmllr.sh: estimating fMLLR transforms a second time.
steps/decode_fmllr.sh: doing a final pass of acoustic rescoring.
steps/diagnostic/analyze_lats.sh --cmd run.pl --mem 8G exp/tri5a/graph exp/tri5a/decode_dev
steps/diagnostic/analyze_lats.sh: see stats in exp/tri5a/decode_dev/log/analyze_alignments.log
Overall, lattice depth (10,50,90-percentile)=(1,2,15) and mean=6.2
steps/diagnostic/analyze_lats.sh: see stats in exp/tri5a/decode_dev/log/analyze_lattice_depth_stats.log
+ steps/score_kaldi.sh --cmd 'run.pl --mem 8G' data/dev exp/tri5a/graph exp/tri5a/decode_dev
steps/score_kaldi.sh --cmd run.pl --mem 8G data/dev exp/tri5a/graph exp/tri5a/decode_dev
steps/score_kaldi.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ steps/scoring/score_kaldi_cer.sh --stage 2 --cmd 'run.pl --mem 8G' data/dev exp/tri5a/graph exp/tri5a/decode_dev
steps/scoring/score_kaldi_cer.sh --stage 2 --cmd run.pl --mem 8G data/dev exp/tri5a/graph exp/tri5a/decode_dev
steps/scoring/score_kaldi_cer.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ echo 'local/score.sh: Done'
local/score.sh: Done
sv@HP:~/lkaldi/egs/aishell/s5$ steps/decode_fmllr.sh --cmd "$decode_cmd" --nj 10 --config conf/decode.config exp/tri5a/graph data/test exp/tri5a/decode_test || exit 1;
steps/decode_fmllr.sh --cmd run.pl --mem 8G --nj 10 --config conf/decode.config exp/tri5a/graph data/test exp/tri5a/decode_test
steps/decode.sh --scoring-opts  --num-threads 1 --skip-scoring false --acwt 0.083333 --nj 10 --cmd run.pl --mem 8G --beam 8.0 --model exp/tri5a/final.alimdl --max-active 2000 exp/tri5a/graph data/test exp/tri5a/decode_test.si
decode.sh: feature type is lda
steps/diagnostic/analyze_lats.sh --cmd run.pl --mem 8G exp/tri5a/graph exp/tri5a/decode_test.si
steps/diagnostic/analyze_lats.sh: see stats in exp/tri5a/decode_test.si/log/analyze_alignments.log
Overall, lattice depth (10,50,90-percentile)=(1,3,16) and mean=6.7
steps/diagnostic/analyze_lats.sh: see stats in exp/tri5a/decode_test.si/log/analyze_lattice_depth_stats.log
+ steps/score_kaldi.sh --cmd 'run.pl --mem 8G' data/test exp/tri5a/graph exp/tri5a/decode_test.si
steps/score_kaldi.sh --cmd run.pl --mem 8G data/test exp/tri5a/graph exp/tri5a/decode_test.si
steps/score_kaldi.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ steps/scoring/score_kaldi_cer.sh --stage 2 --cmd 'run.pl --mem 8G' data/test exp/tri5a/graph exp/tri5a/decode_test.si
steps/scoring/score_kaldi_cer.sh --stage 2 --cmd run.pl --mem 8G data/test exp/tri5a/graph exp/tri5a/decode_test.si
steps/scoring/score_kaldi_cer.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ echo 'local/score.sh: Done'
local/score.sh: Done
steps/decode_fmllr.sh: feature type is lda
steps/decode_fmllr.sh: getting first-pass fMLLR transforms.
steps/decode_fmllr.sh: doing main lattice generation phase
^Z
[1]+  Stopped                 steps/decode_fmllr.sh --cmd "$decode_cmd" --nj 10 --config conf/decode.config exp/tri5a/graph data/test exp/tri5a/decode_test
exit
There are stopped jobs.
sv@HP:~/lkaldi/egs/aishell/s5$ steps/decode_fmllr.sh --cmd "$decode_cmd" --nj 10 --config conf/decode.config exp/tri5a/graph data/test exp/tri5a/decode_test || exit 1;
steps/decode_fmllr.sh --cmd run.pl --mem 8G --nj 10 --config conf/decode.config exp/tri5a/graph data/test exp/tri5a/decode_test
steps/decode.sh --scoring-opts  --num-threads 1 --skip-scoring false --acwt 0.083333 --nj 10 --cmd run.pl --mem 8G --beam 8.0 --model exp/tri5a/final.alimdl --max-active 2000 exp/tri5a/graph data/test exp/tri5a/decode_test.si
decode.sh: feature type is lda
steps/diagnostic/analyze_lats.sh --cmd run.pl --mem 8G exp/tri5a/graph exp/tri5a/decode_test.si
steps/diagnostic/analyze_lats.sh: see stats in exp/tri5a/decode_test.si/log/analyze_alignments.log
Overall, lattice depth (10,50,90-percentile)=(1,3,16) and mean=6.7
steps/diagnostic/analyze_lats.sh: see stats in exp/tri5a/decode_test.si/log/analyze_lattice_depth_stats.log
+ steps/score_kaldi.sh --cmd 'run.pl --mem 8G' data/test exp/tri5a/graph exp/tri5a/decode_test.si
steps/score_kaldi.sh --cmd run.pl --mem 8G data/test exp/tri5a/graph exp/tri5a/decode_test.si
steps/score_kaldi.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ steps/scoring/score_kaldi_cer.sh --stage 2 --cmd 'run.pl --mem 8G' data/test exp/tri5a/graph exp/tri5a/decode_test.si
steps/scoring/score_kaldi_cer.sh --stage 2 --cmd run.pl --mem 8G data/test exp/tri5a/graph exp/tri5a/decode_test.si
steps/scoring/score_kaldi_cer.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ echo 'local/score.sh: Done'
local/score.sh: Done
steps/decode_fmllr.sh: feature type is lda
steps/decode_fmllr.sh: getting first-pass fMLLR transforms.
steps/decode_fmllr.sh: doing main lattice generation phase
steps/decode_fmllr.sh: estimating fMLLR transforms a second time.
steps/decode_fmllr.sh: doing a final pass of acoustic rescoring.
steps/diagnostic/analyze_lats.sh --cmd run.pl --mem 8G exp/tri5a/graph exp/tri5a/decode_test
steps/diagnostic/analyze_lats.sh: see stats in exp/tri5a/decode_test/log/analyze_alignments.log
Overall, lattice depth (10,50,90-percentile)=(1,3,18) and mean=7.3
steps/diagnostic/analyze_lats.sh: see stats in exp/tri5a/decode_test/log/analyze_lattice_depth_stats.log
+ steps/score_kaldi.sh --cmd 'run.pl --mem 8G' data/test exp/tri5a/graph exp/tri5a/decode_test
steps/score_kaldi.sh --cmd run.pl --mem 8G data/test exp/tri5a/graph exp/tri5a/decode_test
steps/score_kaldi.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ steps/scoring/score_kaldi_cer.sh --stage 2 --cmd 'run.pl --mem 8G' data/test exp/tri5a/graph exp/tri5a/decode_test
steps/scoring/score_kaldi_cer.sh --stage 2 --cmd run.pl --mem 8G data/test exp/tri5a/graph exp/tri5a/decode_test
steps/scoring/score_kaldi_cer.sh: scoring with word insertion penalty=0.0,0.5,1.0
+ echo 'local/score.sh: Done'
local/score.sh: Done
sv@HP:~/lkaldi/egs/aishell/s5$ 
sv@HP:~/lkaldi/egs/aishell/s5$ steps/align_fmllr.sh --cmd "$train_cmd" --nj 10 data/train data/lang exp/tri5a exp/tri5a_ali || exit 1;
steps/align_fmllr.sh --cmd run.pl --mem 8G --nj 10 data/train data/lang exp/tri5a exp/tri5a_ali
steps/align_fmllr.sh: feature type is lda
steps/align_fmllr.sh: compiling training graphs
steps/align_fmllr.sh: aligning data in data/train using exp/tri5a/final.alimdl and speaker-independent features.
steps/align_fmllr.sh: computing fMLLR transforms
steps/align_fmllr.sh: doing final alignment.
steps/align_fmllr.sh: done aligning data.
steps/diagnostic/analyze_alignments.sh --cmd run.pl --mem 8G data/lang exp/tri5a_ali
steps/diagnostic/analyze_alignments.sh: see stats in exp/tri5a_ali/log/analyze_alignments.log
121 warnings in exp/tri5a_ali/log/align_pass2.*.log
73 warnings in exp/tri5a_ali/log/align_pass1.*.log

第11部分：nnet3

继续：Kaldi单步完美运行AIShell v1 S5之五：chain DNN
继续：Kaldi单步完美运行AIShell v1 S5之四：nnet3 DNN
回头：Kaldi单步完美运行AIShell v1 S5之三：三音素TriPhone
回头：Kaldi单步完美运行AIShell v1 S5之二：单音素MonoPhone
回头：Kaldi单步完美运行AIShell v1 S5之一：MONO前

其他参考：Kaldi完美运行TIMIT完整结果（含DNN）